次へ: Translation Accuracy of Proposed
上へ: Automatic Evaluation of All
戻る: Simple Sentences
We made 10,000 test sentences that combined 408 sentences for the A-rank and 9,592 sentences for the baseline(Moses).
表 XX:
Comparison of proposed method and baseline(Moses)
|
|
In Table XX, the BLEU score of the "Proposed+Baseline" was higher than the baseline(Moses) by 0.0012. This proves the effectiveness of the "Proposed+Baseline".
平成25年9月17日
![\scalebox{1.0}[1.0]{
\begin{tabular}{\vert l\vert c\vert c\vert}
\hline
& BLEU...
...0999 & 3.1267\\ \hline
Baseline(Moses) & 0.0987 & 3.1100\\ \hline
\end{tabular}}](img30.png)